The Klarna Product-Page Dataset

Stream data with DDA:

from dagshub.streaming import DagsHubFilesystem

fs = DagsHubFilesystem(".", repo_url="https://dagshub.com/DagsHub-Datasets/klarna_productpage_dataset-dataset")

fs.listdir("s3://klarna-research-public-datasets/")

Description:

A collection of 51,701 product pages from 8175 e-commerce websites across 8 markets (US, GB, SE, NL, FI, NO, DE, AT) with 5 manually labelled elements, specifically, the product price, name and image, add-to-cart and go-to-cart buttons. The dataset was collected between 2018 and 2019 and is made available has MHTML and as WebTraversalLibrary-format snapshots.

Contact:

A collection of 51,701 product pages from 8175 e-commerce websites across 8 markets (US, GB, SE, NL, FI, NO, DE, AT) with 5 manually labelled elements, specifically, the product price, name and image, add-to-cart and go-to-cart buttons. The dataset was collected between 2018 and 2019 and is made available has MHTML and as WebTraversalLibrary-format snapshots.

Update Frequency:

The dataset is not expected to update frequently.

Managed By:

Web Automation Research, Klarna

Resources:

1. resource:
   • Description: Bucket containing the two datasets (one in the MHTML and one in the WTL snapshot formats) as tar-balls.
   • ARN: arn:aws:s3:::klarna-research-public-datasets/
   • Region: eu-west-1
   • Type: S3 Bucket

Tags:

internet, natural language processing, computer vision, commerce, deep learning, machine learning, information retrieval, graph